Add initial JID checks

The rexmpp_jid_check function now ensures that JID parts are valid
UTF-8 strings, and that only allowed code points (per RFC 8265) are
used in those. Though there is a few more checks to perform still.

1 files changed, 114 insertions, 1 deletions

diff --git a/src/rexmpp_jid.c b/src/rexmpp_jid.c
index f19bd7c..ee7b326 100644
--- a/src/rexmpp_jid.c
+++ b/src/rexmpp_jid.c
@@ -8,10 +8,14 @@
 
 #include <stddef.h>
 #include <string.h>
+#include <stdio.h>
+#include <unicode/ustring.h>
+#include <unicode/uset.h>
+#include <unicode/uspoof.h>
 #include "rexmpp_jid.h"
 
 int rexmpp_jid_parse (const char *str, struct rexmpp_jid *jid) {
-  const char *resource = NULL, *domain = NULL;
+  const char *resource = NULL, *domain = str;
   size_t i;
   size_t resource_len = 0, local_len = 0;
   size_t domain_len, bare_len, full_len = strlen(str);
@@ -63,3 +67,112 @@ int rexmpp_jid_parse (const char *str, struct rexmpp_jid *jid) {
 
   return 0;
 }
+
+/* <https://tools.ietf.org/html/rfc7622#section-3>,
+   <https://tools.ietf.org/html/rfc8265#section-3.3> */
+int rexmpp_jid_check (struct rexmpp_jid *jid) {
+  UErrorCode err = U_ZERO_ERROR;
+  UChar local[1023], domain[1023], resource[1023];
+  int32_t local_len = 0, domain_len = 0, resource_len = 0;
+
+  /* Initial length checks are performed on parsing. */
+
+  /* Ensure that it's all valid UTF-8. */
+
+  u_strFromUTF8(local, 1023, &local_len, jid->local, -1, &err);
+  if (U_FAILURE(err)) {
+    return 0;
+  }
+  /* TODO: IDNA2008 on domain part. */
+  u_strFromUTF8(domain, 1023, &domain_len, jid->domain, -1, &err);
+  if (U_FAILURE(err)) {
+    return 0;
+  }
+  u_strFromUTF8(resource, 1023, &resource_len, jid->resource, -1, &err);
+  if (U_FAILURE(err)) {
+    return 0;
+  }
+
+  /* TODO: width mapping. */
+
+
+  /* Check character classes */
+  USpoofChecker *sc;
+  int32_t spoof;
+
+  /* IdentifierClass: {Ll, Lu, Lo, Nd, Lm, Mn, Mc} + 0x21 to 0x7e */
+  USet *identifier_chars = uset_openEmpty();
+  uset_applyIntPropertyValue(identifier_chars, UCHAR_GENERAL_CATEGORY_MASK,
+                             U_GC_LL_MASK | U_GC_LU_MASK | U_GC_LO_MASK |
+                             U_GC_ND_MASK | U_GC_LM_MASK | U_GC_MN_MASK |
+                             U_GC_MC_MASK,
+                             &err);
+  if (U_FAILURE(err)) {
+    return 0;
+  }
+  uset_addRange(identifier_chars, 0x21, 0x7e);
+
+  sc = uspoof_open(&err);
+  if (U_FAILURE(err)) {
+    uset_close(identifier_chars);
+    return 0;
+  }
+  uspoof_setChecks(sc, USPOOF_CHAR_LIMIT, &err);
+  if (U_FAILURE(err)) {
+    uset_close(identifier_chars);
+    uspoof_close(sc);
+    return 0;
+  }
+  uspoof_setAllowedChars(sc, identifier_chars, &err);
+  if (U_FAILURE(err)) {
+    uset_close(identifier_chars);
+    uspoof_close(sc);
+    return 0;
+  }
+  spoof = uspoof_check(sc, local, local_len, NULL, &err);
+  if (U_FAILURE(err) || spoof) {
+    uset_close(identifier_chars);
+    uspoof_close(sc);
+    return 0;
+  }
+
+  /* FreeformClass: Zs, {Sm, Sc, Sk, So}, {Pc, Pd, Ps, Pe, Pi, Pf,
+     Po}, toNFKC(cp) != cp ones, {Lt, Nl, No, Me}, and IdentifierClass
+     ones. */
+
+  /* TODO: https://tools.ietf.org/html/rfc8264#section-9.17 */
+  USet *freeform_chars = uset_openEmpty();
+  uset_applyIntPropertyValue(freeform_chars, UCHAR_GENERAL_CATEGORY_MASK,
+                             U_GC_ZS_MASK | U_GC_SM_MASK | U_GC_SC_MASK |
+                             U_GC_SK_MASK | U_GC_SO_MASK | U_GC_PC_MASK |
+                             U_GC_PD_MASK | U_GC_PS_MASK | U_GC_PE_MASK |
+                             U_GC_PI_MASK | U_GC_PF_MASK | U_GC_PO_MASK |
+                             U_GC_LT_MASK | U_GC_NL_MASK | U_GC_NO_MASK |
+                             U_GC_ME_MASK,
+                             &err);
+  if (U_FAILURE(err)) {
+    uset_close(freeform_chars);
+    uset_close(identifier_chars);
+    uspoof_close(sc);
+    return 0;
+  }
+  uset_addAll(freeform_chars, identifier_chars);
+  uset_close(identifier_chars);
+
+  spoof = uspoof_check(sc, resource, resource_len, NULL, &err);
+  if (U_FAILURE(err) || spoof) {
+    uset_close(freeform_chars);
+    uspoof_close(sc);
+    return 0;
+  }
+  uset_close(freeform_chars);
+  uspoof_close(sc);
+
+  /* TODO: case mapping, u_strToLower */
+
+  /* TODO: normalization, unorm2_normalize */
+
+  /* TODO: directionality */
+
+  return 1;
+}